Why AI intake is more important than it sounds

The problem no one talks about

Every contract that enters a legal platform starts in the same place: an intake form.

And for years, that form has been the place where legal asks business users to do something they're not equipped to do, interpret a contract.

"What are the termination rights under this agreement?"

"Is there an indemnity cap? If so, what is it?"

"What governing law applies?"

Reasonable questions for a lawyer. Completely opaque for a procurement manager or marketing director who just wants to get a vendor agreement approved and move on with their day.

So they guess. They leave things blank. Or they bypass the process entirely and send the contract to legal in a Slack message or an email.

The result is a data quality problem that compounds over time. Legal teams end up with a contract repository full of incomplete, inconsistent, or simply wrong key facts. Which makes risk assessment harder, search less useful, and the kind of pattern recognition that good legal ops depends on is basically impossible.

This is a quiet problem. It doesn't create incidents. It just creates noise, friction, and a persistent gap between what legal needs and what it actually gets.

What we built

The core idea is simple: instead of asking the business user to answer questions about the contract, we ask the AI.

When a user uploads a document to the Approve & eSign intake form, Plexus AI reads the contract and automatically fills in the answers, both standard intake questions and any custom questions your legal team has configured for that document type.

The user's job changes from "answer these 15 questions" to "review what AI found."

That's a meaningful shift. It removes the need for legal experience from the intake process entirely. The answers come from the contract itself, not from a business user's best guess at what "force majeure" means.

In our testing, AI fills approximately 75% of custom questions directly from the document. For standard fields like parties, dates, and document title, accuracy is higher. Every AI-suggested answer is editable, users can override any field before submitting. Questions the document can't answer stay blank. The AI doesn't fabricate; it only extracts what it can find.

What customers are saying

The reaction has been clear.

"The autofill saves so much time. Much needed upgrade. Works wonders. Five stars."

"Delighted with this enhancement, was fed up filling details out twice for legal review. Always got cheesed off at the inefficient duplication of effort."

That second one is worth dwelling on. The frustration isn't just about the intake form being slow, it's about being asked to do the same work twice. Once by the legal team before submitting, and once again during final submission. When AI handles the extraction, that duplication disappears.

Cheesed off to five stars. The gap is smaller than we thought.

What this means for legal teams

The most immediate impact is data quality.

Key facts that were previously incomplete, inaccurate, or simply missing now come pre-populated from the source document. Legal teams get more reliable data on every contract that enters the platform, without having to chase business users for corrections.

Over time, that compounds. Better intake data means better search. Better search means faster risk identification. Faster risk identification means the legal team can spend more time on the work that actually matters.

What this means for business users

For business users, the change is more immediate.

A form that used to take 15–20 minutes now takes 2. More importantly, it no longer requires reading and interpreting legal language they were never expected to know.

That reduces friction. Reduced friction means higher adoption. Higher adoption means more contracts going through the legal process, not around it.

The bigger picture

I want to be clear about what this feature is and isn't.

It's not a flagship announcement. It's not going to make headlines. But it's a meaningful step toward something we care about deeply: a legal platform that works as hard for the business as it does for the legal team.

The intake form has always been where those two sides of the business collide and could create friction. AI intake is the beginning of making that collision disappear.

We're building toward a legal operating system where submitting a contract is as easy as uploading a file. Where legal gets the data it needs without burdening the people who generated it. Where the friction between legal and the rest of the business gets smaller with every release.

This feature is a quiet step toward that vision. But it's a real one.

What it’s not and what is coming.

A few things worth being explicit about.

This doesn't replace human review. AI-suggested answers are a starting point, confirmed by the user before submission.

Automatic document type detection is coming. More sophisticated extraction is coming. The goal, longer term, is an intake experience that feels less like filling out a form and more like having someone do it for you.

For now: the form fills itself. The user reviews. Legal gets the data it actually needs.

Don't sleep on this quiet but impactful capability.